JP2004120503A - Data transmission method and apparatus thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a data transmission method for performing data transfer between low-rate data transmission networks via a high-rate data transmission network, and also to provide a data transmission apparatus having a private line and equivalent characteristics thereto. <P>SOLUTION: A multiplexer 2 multiplexes packet data from the low-rate data transmission network having a plurality of ports on byte unit basis and on a time division basis. A packet frame generator/transmitter 3 inserts multiplexing data and a valid/invalid field in an information field of the packet frame to be transmitted to the high-rate data transmission network, applies a header from a header generator 4 to the inserted data, and then transmits it to the high-rate network. A demultiplexer 8 demultiplexes the multiplexed data of the information field in the received packet frame, restores it to the packet data, and transmits the packet data to the other low-rate network when a valid/invalid bit ENB is valid. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a data transmission method and a data transmission apparatus for multiplexing data of a low-rate data transmission network and transmitting the multiplexed data via a high-rate data transmission network.
[0002]
[Prior art]
As a high-rate data transmission network, for example, SONET (Synchronous Optical Network) or SDH (Synchronous Digital Hierarchy) is known. Gigabit Ethernet (registered trademark) such as 1000Base-T is also known. In any case, high-rate data transmission of several Gbps or more is possible. As low-rate data transmission networks, for example, 10Base-T having a transmission rate of 10 Mbps, 100Base-T having a transmission rate of 100 Mbps, and the like are known. A packet over SONET (POS) is known as a system for transmitting data of a low-rate data transmission network via SONET of a high-rate data transmission network.
[0003]
FIG. 6 is an explanatory diagram of a conventional data transmission system, in which 61-1 to 61-n, 62-1 to 62-m are terminal devices, 63 and 64 are multiplexing devices, 65 and 66 are network terminating devices, and 67 is 1 shows a transmission network that performs high-speed transmission by SONET or SDH. Between the terminal devices 61-1 to 61-n and 62-1 to 62-m and the multiplexing devices 63 and 64, for example, the above-described 100Base-T low-rate data transmission network is configured. For example, the above-described 1000Base-T high-rate data transmission network is configured between the network termination devices 65 and 66, and the transmission network 67 configures a high-rate data transmission network based on SONET or SDH.
[0004]
The packet data from the terminal devices 61-1 to 61-n and 62-1 to 62-m is packet-multiplexed by the multiplexers 63 and 64 and transmitted to the network terminating devices 65 and 66. The network terminating devices 65 and 66 map the packet frame obtained by multiplexing the packet data to the payload of the SONET frame or the SDH frame and transmit the packet frame via the transmission network 67. The network terminating devices 65 and 66 demap the packet frame from the payload of the received SONET frame or SDH frame, and transmit the packet frame to the multiplexers 63 and 64 corresponding to the destination. The multiplexing devices 63 and 64 transmit the demultiplexed packet data to the terminal devices 61-1 to 61-n and 62-1 to 62-m corresponding to the destination. Thereby, data transmission can be performed between the terminal devices.
[0005]
When performing packet multiplexing, for example, as shown in FIG. 7, when the number of ports corresponding to the low-rate data transmission network is eight, buffers FiFo1 to FiFo8 corresponding to the ports and a multiplexing unit MUX are provided, and 100Base-T 100Mbps packet data received via the low-rate data transmission network is temporarily stored in buffers Fifo1 to Fifo8, multiplexed by a multiplexing unit MUX in a packet correspondence manner, and as a 1 Gbps packet frame, 1000Base-T high-rate data is transmitted. Send to the transmission network.
[0006]
In this case, for example, the packets A1, A2, A3,... To be input to the buffer FiFo1 are 64 bytes long, and the packets B1, C1, D1, E1, F1, G1, H1, B2 to be input to the buffers FiFo2 to FiFo8, respectively. Packets A1, B1, C1, D1, E1, F1, G1, H1, A2,... Each have a length of 1536 bytes and each packet has a length of 12 bytes. Multiplexed in this order and transmitted to the high-rate data transmission network, the multiplex transmission time of the packets for the seven ports of the buffers Fifo2 to Fifo8 is 86 μs. That is, (1536 + 12) × 7 × 8 × 10 ⁻⁹ = 86 × 10 ⁻⁶ [s].
[0007]
For this reason, the buffer FiFol requires a capacity to accumulate 143 packets of 64 bytes in length during 86 μs, and a delay of 86 μs occurs. In order to absorb such fluctuations, the receiving side also needs a buffer for storing the demultiplexed packets, and the delay caused by this buffer is also added.
[0008]
When a VLAN (Virtual Local Network) or the like is applied, it is necessary to use a multiplexed VLAN or to perform VLAN tag allocation control. In this case, if the low-rate data transmission network and the high-rate data transmission network have different carriers, the VLAN specifications may not be the same, which may make implementation difficult.
[0009]
Also, two high-rate data transmission network packets, for example, 1.25 Gbps packets are converted to a 1 Gbps transmission rate by 8B10B code conversion, and the two systems of packets are multiplexed twice in byte units to achieve a 2.4 Gbps high rate. A system for transmitting data via a rate data transmission network is known (for example, see Patent Document 1).
[0010]
[Patent Document 1]
Japanese Patent Application No. 2001-45069 (FIG. 1)
[0011]
[Problems to be solved by the invention]
When multiplexing packet data of a low-rate data transmission network to which a terminal device is connected and transmitting the multiplexed data through a high-rate data transmission network, packet multiplexing is applied, as described with reference to FIG. Fluctuations increase, and a buffer is provided to absorb the fluctuations. However, a delay due to the fluctuations is added, and there is a problem that transmission delay increases. Further, when the packet length is long, the time for one transmission is long, which may adversely affect the start of transmission of a packet from another line.
[0012]
For packet multiplexing, for example, it is conceivable to apply time-division multiplexing as disclosed in Patent Document 1 mentioned above to transmit packets from a low-rate data transmission network to a high-rate data transmission network. In this case, in the low-rate data transmission network, packets are not always transmitted. Therefore, when time-division multiplexing is performed, invalid data is multiplexed, and there is a problem that it is difficult to perform accurate demultiplexing on the receiving side.
[0013]
The present invention solves the conventional problems by enabling fixed bandwidth allocation to a low-rate data transmission network and facilitating identification of valid data or invalid data on the demultiplexing side. Aim.
[0014]
[Means for Solving the Problems]
The data transmission method of the present invention is a data transmission method for performing data transmission between low-rate data transmission networks of about 100 Mbps or less through a high-rate data transmission network of 1 Gbps or more, and a low-speed data transmission method supporting a plurality of ports. The packet data input from the rate data transmission network is time-division multiplexed, a valid / invalid field indicating validity / invalidity of packet data corresponding to a plurality of ports and multiplexed data are inserted into an information field, and a header including a destination address is inserted. Adding the packet frame to the high-rate data transmission network, and demultiplexing and multiplexing the multiplexed data of the information field of the packet frame received via the high-rate data transmission network. A packet using the multiplexed / demultiplexed data corresponding to the port validated by the / invalid field Restore over data, it is intended to include a process of sending the packet data to a port corresponding low-rate data transmission network.
[0015]
In addition, the bandwidth of the low-rate data transmission network corresponding to a plurality of ports is set to, for example, 100 Mbps in total, and a bandwidth obtained by multiplying this bandwidth by the number of ports, for example, a bandwidth of 1 Gbps in the case of eight ports is secured in the high-rate data transmission network. The method includes a step of transmitting data between low-rate data transmission networks via a rate data transmission network.
[0016]
Referring to FIG. 1, the data transmission apparatus of the present invention is a data transmission apparatus for performing data transmission between low-rate data transmission networks via a high-rate data transmission network. A plurality of ports for receiving packet data from the low-rate data transmission network; input buffers 1-1 to 1-8 corresponding to the plurality of ports; A multiplexing unit 2 for performing time-division multiplexing at a predetermined position every time, and adding information to the multiplexed data by the multiplexing unit 2 to indicate a valid / invalid field of packet data corresponding to a plurality of ports. A packet frame generating / sending unit 3 which inserts the packet into a field and adds a header from the header generating unit 4 to form and send a packet frame. A data extraction unit 7 for extracting a valid / invalid field and multiplexed data from an information field of a packet frame received through the multiplexing unit, and demultiplexes the multiplexed data extracted by the data extraction unit 7 for a plurality of ports. The demultiplexer 8 performs the demultiplexing and the output buffers 10-1 to 10-8 corresponding to a plurality of ports for transmitting data indicating validity by a valid / invalid field of the data corresponding to a plurality of ports to the low-rate data transmission network as packet data. Have
[0017]
Also provided are frame check units 9-1 to 9-8 that check packet data that is separated by the multiplex separation unit 8 and that is made valid by the valid / invalid field. Also, the plurality of ports are set to 8 ports, the packet data corresponding to 8 ports is time-division multiplexed in byte units by the multiplexing unit 2, and the valid / invalid field is configured to be 8 bits corresponding to 8 ports. The validity / invalidity of the packet data corresponding to eight ports can be indicated by the bits.
[0018]
BEST MODE FOR CARRYING OUT THE INVENTION
FIG. 1 is an explanatory diagram of an embodiment of the present invention, showing a data transmission apparatus having a configuration in which a low-rate data transmission network side has eight ports and a high-rate data transmission network side has one port. Is an input buffer, 2 is a multiplexing unit (MUX), 3 is a packet frame generation and transmission unit, 4 is a header generation unit, 5 is a packet frame transmission timing generation unit, 6 is a packet frame check unit, 7 is a data extraction unit, 8 Denotes a demultiplexing unit (DEMUX), 9-1 to 9-8 denote frame check units, and 10-1 to 10-8 denote output buffers. ENB indicates a valid / invalid bit.
[0019]
The transmission unit is configured by a configuration including the input buffers 1-1 to 1-8, the multiplexing unit 2, the packet frame generation and transmission unit 3, the header generation unit 4, and the packet frame transmission timing generation unit 5, A receiving unit is configured by a configuration including the frame checking unit 6, the data extracting unit 7, the demultiplexing unit 8, the frame checking units 9-1 to 9-8, and the output buffers 10-1 to 10-8. I have.
[0020]
The packet data input from the low-rate data transmission network is temporarily stored in the input buffers 1-1 to 1-8, respectively, and the multiplexing unit 2 performs time division multiplexing in byte units to generate and transmit packet frames. Transfer to section 3. In the input buffer 1-1 to 1-8 or the multiplexing unit 2, the presence / absence of packet data corresponding to the port is identified, and if there is packet data, a valid / invalid bit ENB indicating validity is indicated. It is input to the packet frame generation / transmission unit 3. Also, the header generation unit 4 generates a header including a transmission destination address for the data transmission device opposed via the high-rate data transmission network and a transmission source address indicating the own data transmission device. input.
[0021]
The packet frame generating / sending unit 3 time-division multiplexes the packet data corresponding to the port in the multiplexing unit 2 in byte units, adds a valid / invalid field by a valid / invalid bit ENB to the head, and inserts it into the information field. Then, a frame check sequence FCS for the data inserted in this information field is obtained and added to the tail, and a header from the header generation unit 4 is added to the head to generate a packet frame. The packet frame is transmitted to the high-rate data transmission network according to the transmission timing signal.
[0022]
When a packet frame is received via the high-rate data transmission network, the packet frame check unit 6 identifies whether or not the packet is addressed to the own data transmission device by a header, and extracts the packet frame addressed to the own data transmission device from the data extraction unit 7. Transfer to The data extraction unit 7 transfers the valid / invalid bit ENB corresponding to the port in the valid / invalid field to the frame check units 9-1 to 9-8 corresponding to the port, or demultiplexes the time-division multiplexed data. Transfer to section 8.
[0023]
The demultiplexing unit 8 demultiplexes the data in byte units and transfers the demultiplexed data to the frame check units 91 to 9-8. The frame check units 9-1 to 9-8 do not perform a check process when the valid / invalid bit ENB indicates invalid. However, when the valid / invalid bit ENB indicates valid, the frame check sequence FCS added to the packet data corresponding to the port. Is used to determine the presence or absence of a transmission error. If there is no error, the data is transmitted to the low-rate data transmission network corresponding to the port via the output buffers 10-1 to 10-8.
[0024]
2A and 2B are explanatory diagrams of the multiplexing process. FIG. 2A shows packet data from the low-rate data transmission network, FIG. 2B shows data P0 in byte units of port 0, and FIG. And (d) show a packet frame. Also, in the packet data of (a), DA is a destination address (Destination Address), SA is a source address (Source Address), T / L is a type and length field, IF is an information field (Information Field), and FCS is a frame check. 5 shows a sequence (Frame Check Sequence).
[0025]
(B) shows the packet data shown in (a) of port 0 among ports 0 to 7 divided into 8 bits, that is, in units of bytes, and shown as P0. (C) shows the byte unit data of the port 0 as P0 and the byte unit data of the ports 1 to 7 as P1 to P7. For example, the data is input to the input buffers 1-1 to 1-7 in FIG. The byte unit data P0 to P7 corresponding to the ports 0 to 7 of the packet data are time-division multiplexed in the multiplexing unit 2, and the head of the input buffer 1-1 to 1-7 or the port from the multiplexing unit 2 A valid / invalid field EN composed of valid / invalid bits ENB corresponding to 0 to 7 is added.
[0026]
In (c), each byte is not shown with the same length for the sake of convenience, but is actually the same length of one byte including the valid / invalid field EN. The valid / invalid bit ENB can indicate "1" as valid and "0" as invalid to indicate valid / invalid. Accordingly, the valid / invalid field EN at the beginning of (c) indicates a case where all are "1", and thus all the packet data of the ports 0 to 7 are valid.
[0027]
(D) shows a packet frame to be transmitted to the high-rate data transmission network, where DA is the destination address in the high-rate data transmission network, SA is the source address, T / L is the type and length field, and IF is IF. The information field, FCS, indicates a frame check sequence. The multiplexed data in byte units shown in (c) is inserted into this information field IF, and the frame check sequence FCS is calculated and added. A header including a destination address DA, a source address SA, a type and a length field T / L is generated in a header generation unit 4 (see FIG. 1), and the header is added in a packet frame generation / transmission unit 3. Then, a packet frame shown in (d) is generated and transmitted to the high-rate data transmission network. Therefore, this corresponds to a case where packet data of the low-rate data transmission network is time-division multiplexed and encapsulated by a packet frame in the high-rate data transmission network. This makes it possible to directly transfer a MAC layer control packet and the like between terminal devices connected to the low-rate data transmission network.
[0028]
3A and 3B are explanatory diagrams of the multiplexing process. FIG. 3A shows packet data from a low-rate data transmission network of ports 0 to 7, and FIG. 3B shows a packet frame transmitted to a high-rate data transmission network by time division multiplexing. Show. .., P71, P72,... Indicate packet data corresponding to ports 0-7. Assuming that the low-rate data transmission network corresponding to ports 0 to 7 is 100 Mbps, multiplexing for 8 ports requires a transmission bandwidth of 800 Mbps. Therefore, for example, a bandwidth of 1 Gbps is allocated to the high-rate data transmission network. When such bandwidth allocation is performed, each of the low-rate data transmission networks that perform data transmission via the high-rate data transmission network can reliably secure a 100 Mbps bandwidth and have a sufficient packet frame interval. it can. Therefore, there is no need for a buffer capacity for waiting when packets of other ports are long, and there is almost no fluctuation in transmission delay, which is equivalent to a case where low-rate data transmission networks are connected by a dedicated line.
[0029]
Also, in FIG. 3A, even if the timing of the packet data input to the input buffers 1-1 to 1-7 (see FIG. 1) is different, the packet data corresponding to the ports 0 to 7 is multiplexed. In the section 2, time division multiplexing is performed, and a valid / invalid field EN is added every 8 bytes to make the length 9 bytes. If there is no received packet data in the input buffers 1-1 to 1-7 at the time division multiplex timing, the valid / invalid bit ENB is set to an invalid bit of "0" and the received packets are stored in the input buffers 1-1 to 1-7. If there is data, the valid / invalid bit ENB is set to a valid bit of “1”.
[0030]
This is inserted into the information field IF by inserting one or a plurality thereof according to the byte length of the information field IF of the packet frame to be transmitted to the high-rate data transmission network shown in FIG. And sends it out to the high-rate data transmission network. In this case, since the frame becomes a fixed-length frame and is transmitted at a predetermined frame interval, the packet frame is transmitted to the high-rate data transmission network in a fixed cycle, and high-speed transmission can be performed with the transmission band secured in advance.
[0031]
FIG. 4 is an explanatory diagram of the demultiplexing process, and shows the operation of the demultiplexing in the configuration including the data extracting unit 7 and the demultiplexing unit 8 of the receiving unit in FIG. FIG. 3A shows a packet frame received via a high-rate data transmission network, FIG. 3B shows multiplexed data extracted from an information field IF of the packet frame, and FIG. 3C shows demultiplexed data corresponding to port 0. , (D) shows valid data, and (e) shows restored packet data.
[0032]
The packet frame shown in (a) received from the high-rate data transmission network is received by the packet frame check unit 6 (see FIG. 1), and its destination address DA identifies that the packet is addressed to its own data transmission apparatus. Transfer to the extraction unit 7. The data extracting unit 7 extracts multiplexed data from the information field IF and transfers the multiplexed data to the demultiplexing unit 8. The demultiplexing unit 8 or the data extracting unit 7 identifies valid / invalid bits ENB corresponding to ports 0 to 7 of the valid / invalid field EN, and notifies the frame check units 9-1 to 9-8. FIG. 1 shows a case where the valid / invalid bit ENB identified in the data extraction unit 7 is notified to the frame check units 9-1 to 9-8.
[0033]
Also, since the eight valid / invalid bits of the valid / invalid field EN shown in FIG. 4B are all shown as "1", it is shown that the data corresponding to ports 0 to 7 is valid. I have. The demultiplexing unit 8 performs demultiplexing for each of the ports 0 to 7. Therefore, for the port 0, only the data P0 is separated as shown in FIG.
[0034]
In this case, the data P0 indicated as EN = 0 is invalid, and the data P0 indicated as EN = 1 is valid. Therefore, the data P0 in the period TE of EN = 1 is checked, and the data P0 is output to the output buffer 10-1. Stores valid data shown in (d). This data is packet data including a destination address DA, a source address SA, a type / length field T / L, an information field IF, and a frame check sequence FCS, as shown in FIG. To send to.
[0035]
FIG. 5 is an explanatory diagram of a data transmission system, where 31 and 32 are data transmission devices, 33 is a data transmission network, and 34 and 35 are terminal devices. The data transmission devices 31 and 32 have the configuration shown in FIG. 1, and the terminal devices 34 and 35 are connected to the data transmission devices 31 and 32 via a low-rate data transmission network. 32 are connected via a data transmission network 33, and the data transmission devices 31 and 32 time-division multiplex the packet data received via the low-rate data transmission network as described above, and enable / disable the port correspondence. A valid / invalid field including a bit is added and transmitted to the data transmission network 33.
[0036]
The data transmission devices 31 and 32, which have received the packet frame via the data transmission network 33, demultiplex and refer to the valid / invalid field and use only the valid data to transmit the terminal device via the low-rate data transmission network. 34 and 35. Therefore, the data transmission network 33 constituting the high-rate data transmission network is equivalent to a case where a dedicated line connection is made between the low-rate data transmission networks as indicated by a chain line, and the low-rate data transmission network has Various control information and the like can be transmitted without performing protocol conversion or the like, similarly to the case of the dedicated line.
[0037]
The present invention is not limited to each of the above-described embodiments, and various additions and changes are possible. If the above-described plurality of ports is, for example, 16, the valid / invalid field has a 16-bit configuration. By performing multiplexing and demultiplexing in units of 2 bytes, packet data from the low-rate data transmission network can be multiplexed and transmitted via the high-rate data transmission network.
[0038]
As described above, the packet data of the low-rate data transmission network is encapsulated from the packet frame of the high-rate data transmission network. Therefore, control packets and the like in the low-rate data transmission network are encapsulated. Since transmission is possible, it is possible to completely separate the network management between the high-rate data transmission network and the low-rate data transmission network. Thereby, even in the case of a high-rate data transmission network and a low-rate data transmission network of different carriers, a VLAN tag can be added, and multiplexing of a higher-order multiplexed VLAN or the like is possible.
[0039]
【The invention's effect】
As described above, according to the present invention, packet data corresponding to a plurality of ports is time-division multiplexed, and a valid / invalid field EN indicating validity / invalidity of the packet data corresponding to a plurality of ports is added to the byte multiplexed data. To form a packet frame by adding a header including a destination address and the like in the high-rate data transmission network, send the packet frame to the high-rate data transmission network, receive this packet frame, and It demultiplexes the extracted multiplexed data and processes only the data indicating validity by referring to the valid / invalid field to restore the packet data to be transmitted to the low-rate data transmission network. By using a fixed bandwidth for low-rate data transmission when performing multiplex transmission via a high-rate data transmission network. The networks can provide a dedicated line equivalent data transmission environments. Therefore, there is an advantage that it is possible to omit a large-capacity buffer waiting for transmission to reduce the cost and to reduce transmission delay fluctuation and transmission delay, thereby enabling data transmission with high transmission efficiency.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram of an embodiment of the present invention.
FIG. 2 is an explanatory diagram of a multiplexing process.
FIG. 3 is an explanatory diagram of a multiplexing process.
FIG. 4 is an explanatory diagram of a demultiplexing process.
FIG. 5 is an explanatory diagram of a data transmission system.
FIG. 6 is an explanatory diagram of a conventional data transmission system.
FIG. 7 is an explanatory diagram of conventional multiplexed data transmission.
[Explanation of symbols]
1-1 to 1-8 Input Buffer 2 Multiplexer (MUX)
3 packet frame generation / transmission unit 4 header generation unit 5 packet frame transmission timing generation unit 6 packet frame check unit 7 data extraction unit 8 demultiplexing unit (DEMUX)
9-1 to 9-8 Frame Check Unit 10-1 to 10-8 Output Buffer

Claims (5)

In a data transmission method for transmitting data between low-rate data transmission networks via a high-rate data transmission network,
The packet data input from the low-rate data transmission network corresponding to a plurality of ports is time-division multiplexed, and a valid / invalid field indicating validity / invalidity of the packet data corresponding to the plurality of ports and multiplexed data are inserted into an information field. Composing a packet frame by adding a header including a destination address, and transmitting the packet frame to the high-rate data transmission network;
Demultiplexing the multiplexed data of the information field of the packet frame received via the high-rate data transmission network, and restoring the packet data with the demultiplexed data corresponding to the port validated by the valid / invalid field; Transmitting the packet data to the low-rate data transmission network corresponding to the port. The bandwidths of the low-rate data transmission network corresponding to the plurality of ports are all the same, and a bandwidth obtained by multiplying the bandwidth by the number of ports is secured at least in the high-rate data transmission network. 2. The data transmission method according to claim 1, further comprising the step of transmitting data between said low-rate data transmission networks via said pre-assigned high-rate data transmission network. In a data transmission device for performing data transmission between low-rate data transmission networks via a high-rate data transmission network,
Having a transmitting unit and a receiving unit for the high-rate data transmission network,
The transmitting unit includes a plurality of ports for receiving packet data from the low-rate data transmission network, an input buffer corresponding to the plurality of ports, a multiplexing unit for time-division multiplexing the packet data, and multiplexing by the multiplexing unit. A valid / invalid field indicating valid / invalid of the packet data corresponding to the plurality of ports is added to the encoded data, inserted into the information field, and a header from a header generation unit is added to form a packet frame and transmitted. A frame generation and transmission unit,
A receiving unit configured to extract the valid / invalid field and the multiplexed data from an information field of a packet frame received via the high-rate data transmission network; and a multiplexing unit configured to extract the multiplexed data extracted by the data extracting unit. A demultiplexing unit that performs time division multiplexing / demultiplexing of data corresponding to the plurality of ports; and a plurality of units that transmits data indicating validity by the valid / invalid field of the data corresponding to the plurality of ports to the low-rate data transmission network as packet data. A data transmission device, comprising: an output buffer corresponding to a port. 4. The data transmission apparatus according to claim 3, further comprising a frame check unit that checks packet data that is separated by the demultiplexing unit and is valid according to the valid / invalid field. The plurality of ports are set to eight ports, the multiplexing unit multiplexes the packet data corresponding to eight ports in byte units, and sets the valid / invalid field to an 8-bit configuration corresponding to the eight ports. 4. The data transmission apparatus according to claim 3, further comprising a configuration for displaying validity / invalidity of the packet data corresponding to the eight ports using bits, wherein a multiplexing position is assigned in advance for each packet data.

Images